intertwingly

Ironically, this makes Dave's feed incompatible with his spec. It doesn't matter to namespace-ignorant aggregators, but it *does* matter to aggregators that understand namespaces.

Sam, there are fundamental differences that simple module movement won't fix. One is ownership of RSS -- who owns this spec? We can't keep going through this with every release.

The other is the basic underlying technology of RSS. If RDF remains in this picture, it needs to be acknowledged as part of the underlying design rather than sitting like the elephant in the corner, which we pretend to ignore if it makes acceptance easier.

If RDF needs to be proved, then let us prove it. But trying to merge two disparate RSS specs based on module movement is, bluntly, not workable. It would only lead to continued confusion.

Shelley, by all means, prove RDF.

As to ownership, my hope is that we can come to agreement on a simple interoperable core that everybody agrees upon, and then individual namespaces owned by their authors.

Prove RDF. How many apps and specs and effort will it take to prove it? Never mind, I'll see what I can whip up in the next hour or two.

As for 'we', who's 'we'? The RSS-Dev team and Dave Winer? You? Me? Who constitutes an acceptable 'we' that the community of RSS users would accept 'we' as a representative group?

I'm not disparaging your attempts to find a solution to the diverged specifications. But quick solutions (here's the core, here's namespace, we'll find someway to munge this all together and be one happy family, and all you people who find RDF to be important, well you haven't proved anything to me or to Dave, so until you do, we'll just continue on and work it out, somehow, in the future) are what got us into this mess in the first place.

With namespaces agreed upon in principle, cleaning up the "core," to me, is the
most important to simplify and uniting the format. So I agree with Sam's notion of removing (depreciating?) redundent and "optional" tags and moving them to modules.

This said, I understand and appreciate burningbird's point that this does not entirely settle things. RDF was a design goal of 0.90 and 1.0. Its merit, like these tags we'd like to see removed from core, needs to be considered and debated seperately and in the same spirit.

I'm trying to keep an open mind for peace (and progress) here.

Shelley, Timothy:

Baby steps. The way to break a logjam is one log at a time.

Ray Ozzie today describes how the US Constitution's authors took into account the need for a somewhat-decentralized committee (Congress), a centralizing power (President), and an arbiter (Supreme Court). Seems like rss-dev is the equivalent of Congress, Dave's trying to be the President, and you, dear Sam, are assuming black robes. Bless you, sir.

Mike, don't blow it up out of proportion. RSS is a format for syndicating Web content. If you get zen-like instead of Jefferson-like, you can ask it what it is, and guess what it can tell you. Sam did a survey a couple of days ago listing the frequency of each of the elements in recorded feeds on Syndic8 (understanding that Syndic8 is imperfect, but the best source we have). The President, Congress and Supreme Court is the installed base. That's how it works on formats that are going on five years old. You guys all talk like there something you can do to change this thing called RSS. You can't, I can't, no one can.

Dave, you can and are influencing RSS. Even those that don’t agree with you are taking the introduction of namespaces into Userland’s RSS as a positive step and have produced overtures of their own. With consensus, RSS 2.0 has the potential to unite both RSS camps and create a single, simple syndication format that can be extended to syndicate almost any content. Two major competing syndication formats hurts everybody no matter what they are called and the RSS name has momentum.

How many namespaces are there going to be for what are essentially the same tags?
as ziv said "it *does* matter to aggregators that understand namespaces."

and I'm betting that as time goes on that will be more and more aggregators.

If the technology used for working with the namespaced tree is one developed for working with xml, for example xslt, dom, sax etc. then having namespaces declared willy-nilly breaks rather than improves rss.

Personally I thought RSS 1.0 was an example of what has been termed neurotic namespacing(by Joe English), and hence, despite the big names who were involved with it, bad design. This is of course my own personal take, I tend to disdain complexity.
Since Userland seems to be good at avoiding the needless complexities that most specifications tend to force on us I had thought their approach was as it was partly because they wished to avoid namespaces and the headaches that they can cause. Life just got a lot more difficult for me, in that the development I will have to do in the next couple months will involve writing an aggregator.

Neurotic namespaces, as defined by Joe English, refers to "[mapping] the same namespace prefix to two different namespace URIs at different points."

Per Joe's definitions, RSS 1.0 is not only "sane", but "normal" when following recommended practice.

Andrew Grumet and I finished debugging  this problem yesterday.  It took a number of sessions, via AIM, phone, and email, to finally track it down.... [more]

Trackback from Sam Ruby